Bug 34898

Summary: REGRESSION(54694): websocket/tests/bufferedAmount-after-close.html failed on Leopard Commit Bot
Product: WebKit Reporter: Eric Seidel (no email) <eric>
Component: Tools / TestsAssignee: Nobody <webkit-unassigned>
Severity: Normal CC: ap, ukai
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: PC   
OS: OS X 10.5   
Bug Depends on:    
Bug Blocks: 34633, 34702, 34822, 34829    
Description Flags
crash log from commit bot none

Description Eric Seidel (no email) 2010-02-12 09:52:41 PST
websocket/tests/bufferedAmount-after-close.html failed on Leopard Commit Bot

The commit bot currently: updates, builds, runs all the tests once, if the tests pass then it: applies the patch, builds and runs the tests again.

So when the commit-bot sees a failure, we know its a flakey test.  It saw websocket/tests/bufferedAmount-after-close.html fail:

This is a very recent test:
Comment 1 Eric Seidel (no email) 2010-02-12 14:35:24 PST
The commit queue saw this again!

I think something must be wrong with the code if it's crashing.  I can get a crash log for you.
Comment 2 Eric Seidel (no email) 2010-02-12 14:38:04 PST
Created attachment 48665 [details]
crash log from commit bot
Comment 3 Eric Seidel (no email) 2010-02-12 14:39:50 PST
I actually have 10 of the same crash log from teh commit-queue the first one is from:
Date/Time:       2010-02-11 23:02:33.451 -0800

This is clearly a regression from http://trac.webkit.org/changeset/54694.  I recommend we roll it out.
Comment 4 Eric Seidel (no email) 2010-02-12 14:49:27 PST
Actually the commit-queue has seen this crash 12 times since this was landed. :)  I can make it do a full rebuild, but I suspect the crash is real.
Comment 5 Eric Seidel (no email) 2010-02-12 15:39:19 PST
Rolled out the change.  Duping this against the original bug.

*** This bug has been marked as a duplicate of bug 34633 ***